INDEX

Explanations

future intentions or possibilities expressed by the phrase "will" or "would."

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-0.70

said

-0.62

arthed

-0.61

 adolesc

-0.60

told

-0.59

raq

-0.57

Lear

-0.57

 guessing

-0.55

 Strategy

-0.55

 Provided

-0.55

POSITIVE LOGITS

 someday

1.24

be

1.19

 eventually

1.09

 succeed

1.00

 somehow

0.96

 prevail

0.95

 become

0.93

 emerge

0.92

soon

0.92

 arrive

0.91

Activations Density 0.181%

future intentions or possibilities expressed by the phrase "will" or "would."

No Comments

No Known Activations